177 lines
5.1 KiB
JSON
177 lines
5.1 KiB
JSON
{
|
|
"dossier_dir": "/root/.hermes/wizard-outputs/sphinx/mm-vla-dossier-2026-10-22",
|
|
"generated_utc": "2026-10-22T17:13:24Z",
|
|
"query": "cat:cs.CV AND abs:\"multimodal\"",
|
|
"atom_url": "https://export.arxiv.org/api/query?search_query=cat:cs.CV+AND+abs:%22multimodal%22&start=0&max_results=120&sortBy=submittedDate&sortOrder=descending",
|
|
"atom_http": 200,
|
|
"atom_sha256": "17767bfb4ec4c02fdc21e79d7bf7f96035610ee37356a1edc2a3319a7e4c95be",
|
|
"atom_bytes": 311552,
|
|
"atom_totalResults": 15740,
|
|
"papers": [
|
|
{
|
|
"arxiv_id": "2610.30108v1",
|
|
"title": "Multimodal Embodied Agents: A Unified Capability-centric Perspective from Computer-Use to Robot-Use",
|
|
"primary_category": "cs.RO",
|
|
"categories": [
|
|
"cs.RO",
|
|
"cs.AI",
|
|
"cs.CV",
|
|
"cs.MA",
|
|
"cs.MM"
|
|
],
|
|
"authors": [
|
|
"Yanzhe Chen",
|
|
"Ziyi Yang",
|
|
"Jifeng Zhu",
|
|
"Qiming Huang",
|
|
"Ruihe An",
|
|
"Peiyao Xu",
|
|
"Hesen Yang",
|
|
"Runda Liu",
|
|
"Chang Gong",
|
|
"Zhijun Cao",
|
|
"Zechen Bai",
|
|
"Wenzheng Zeng",
|
|
"Yiqi Lin",
|
|
"Guoqiang Liang",
|
|
"Kevin Yuchen Ma",
|
|
"Kevin Qinghong Lin",
|
|
"Mike Zheng Shou"
|
|
],
|
|
"published": "2026-10-21T17:52:48Z",
|
|
"updated": "2026-10-21T17:52:48Z",
|
|
"html_sha256": "07af14c4b4e8491d61b4940433d03836a02b486ba2d473a360bef9ace8bc3d1f",
|
|
"html_bytes": 1102349,
|
|
"tables": 4,
|
|
"figures": 12,
|
|
"images": 10,
|
|
"dossier": "dossier.json"
|
|
},
|
|
{
|
|
"arxiv_id": "2610.30089v1",
|
|
"title": "ProactEgoMem: Benchmarking Proactive Memory in Streaming Egocentric Video",
|
|
"primary_category": "cs.CV",
|
|
"categories": [
|
|
"cs.CV"
|
|
],
|
|
"authors": [
|
|
"Ruotong Liao",
|
|
"Meikel Kokowski",
|
|
"Xiaoxiao Sun",
|
|
"Yongqin Xian",
|
|
"Laura Bravo S\u00e1nchez",
|
|
"Shiye Su",
|
|
"Federico Tombari",
|
|
"Serena Yeung-Levy",
|
|
"Volker Tresp"
|
|
],
|
|
"published": "2026-10-21T17:37:58Z",
|
|
"updated": "2026-10-21T17:37:58Z",
|
|
"html_sha256": "50947ff1e7c8be12fa01771ec1a2c10efe54842661c19538bb5babdd045107c8",
|
|
"html_bytes": 1879293,
|
|
"tables": 12,
|
|
"figures": 30,
|
|
"images": 12,
|
|
"dossier": "dossier.json"
|
|
},
|
|
{
|
|
"arxiv_id": "2610.29913v1",
|
|
"title": "Asking a Multimodal LLM Whether Someone Fell: Zero-Shot Screening of Low-Resolution Thermal Video in a Laboratory and an Occupied Home",
|
|
"primary_category": "cs.CV",
|
|
"categories": [
|
|
"cs.CV",
|
|
"cs.AI"
|
|
],
|
|
"authors": [
|
|
"Marcos Junca",
|
|
"Mauro Junca"
|
|
],
|
|
"published": "2026-10-21T15:53:17Z",
|
|
"updated": "2026-10-21T15:53:17Z",
|
|
"html_sha256": "ff4f590f25c54fbcc5b32c35fefbaa9aa97fdb81f1fe0601a9287ccd89be81e7",
|
|
"html_bytes": 239906,
|
|
"tables": 6,
|
|
"figures": 12,
|
|
"images": 1,
|
|
"dossier": "dossier.json"
|
|
}
|
|
],
|
|
"files": {
|
|
"REPORT.md": {
|
|
"sha256": "4def6c59fb407158ce03cdbe9851fea0bf60b437bc60d584fceda40d212db9b5",
|
|
"bytes": 6619
|
|
},
|
|
"coverage_probe.json": {
|
|
"sha256": "a52323156933264c56c168084765b49e08c16a3b98ae689a87bfd1c6c6ebbafb",
|
|
"bytes": 28464
|
|
},
|
|
"coverage_probe.py": {
|
|
"sha256": "65916084d57d6a68b3767948735322f6ba0e93ca1d4eb0f0158ee500d5b5a7e2",
|
|
"bytes": 2470
|
|
},
|
|
"2610.30108v1/dossier.json": {
|
|
"sha256": "3426a4710481151dd98c06ac306128a72a481f445b7d4370182673b2531e658d",
|
|
"bytes": 3523
|
|
},
|
|
"2610.30108v1/figures.json": {
|
|
"sha256": "945abcbc1508198dc5b6eeff5ed6cf66ec18437294a5ad6671f1759153356eb5",
|
|
"bytes": 5817
|
|
},
|
|
"2610.30108v1/paper.html": {
|
|
"sha256": "07af14c4b4e8491d61b4940433d03836a02b486ba2d473a360bef9ace8bc3d1f",
|
|
"bytes": 1102349
|
|
},
|
|
"2610.30108v1/paper.txt": {
|
|
"sha256": "be1522c76f455b2872dcc3994de7e6c237f513e85e84ab5a7a36e0d385533471",
|
|
"bytes": 199028
|
|
},
|
|
"2610.30108v1/tables.json": {
|
|
"sha256": "2ecea4a9c8e97761e25cf1bfc0598c66b2c2080985be61bc1f4461ec763bf471",
|
|
"bytes": 14211
|
|
},
|
|
"2610.29913v1/dossier.json": {
|
|
"sha256": "c8fc594eb36367e4b8cd5b675db5dd094cda70c0e5aeea4237ce18a0f5761239",
|
|
"bytes": 3409
|
|
},
|
|
"2610.29913v1/figures.json": {
|
|
"sha256": "b02c21c8a7281772b60c282d6d9e77c1840fbebf9b29661b9c6164ef522c75b9",
|
|
"bytes": 5355
|
|
},
|
|
"2610.29913v1/paper.html": {
|
|
"sha256": "ff4f590f25c54fbcc5b32c35fefbaa9aa97fdb81f1fe0601a9287ccd89be81e7",
|
|
"bytes": 239906
|
|
},
|
|
"2610.29913v1/paper.txt": {
|
|
"sha256": "26fe614558e972064a5d15d79383d16d3f1b1abef9846dc5e6881554e443d5ff",
|
|
"bytes": 74930
|
|
},
|
|
"2610.29913v1/tables.json": {
|
|
"sha256": "0dc9286e5a581416ca874ca5fc61e5e6a518969c1ea77946c01c5638d642fa86",
|
|
"bytes": 7429
|
|
},
|
|
"2610.30089v1/dossier.json": {
|
|
"sha256": "d1f7ab9a02372aec2971757cd336199815452aa75fdbcd72d5a52e1d185a9d90",
|
|
"bytes": 3919
|
|
},
|
|
"2610.30089v1/figures.json": {
|
|
"sha256": "a4faac626d547d0ddba399e0f494d2bd1fa833c7f5c32c0fff7feca7f14783c9",
|
|
"bytes": 10123
|
|
},
|
|
"2610.30089v1/paper.html": {
|
|
"sha256": "50947ff1e7c8be12fa01771ec1a2c10efe54842661c19538bb5babdd045107c8",
|
|
"bytes": 1879293
|
|
},
|
|
"2610.30089v1/paper.txt": {
|
|
"sha256": "ddf9b5c076a157fdff66cc54b03097b2200b315403e548e33bf7639fdec734c6",
|
|
"bytes": 121203
|
|
},
|
|
"2610.30089v1/tables.json": {
|
|
"sha256": "2a7f2756c4609df24fa7000642c4ed00128ae2940474fdc22475a5605f4c2886",
|
|
"bytes": 29693
|
|
},
|
|
"raw/atom_feed.xml": {
|
|
"sha256": "17767bfb4ec4c02fdc21e79d7bf7f96035610ee37356a1edc2a3319a7e4c95be",
|
|
"bytes": 311552
|
|
}
|
|
}
|
|
} |